Researchers at the Shiley Eye Institute, University of California, San Diego, has reported results of AI (artificial intelligence) methods providing significantly accurate outcomes over manual applications on overlaying multimodal retinal imaging in RP patients. According to the researchers, the results of their study “is the first to investigate the efficacy of AI to overlay structural over functional retinal images in patients with RP compared with manual alignment. This constitutes an important step towards developing efficient algorithms for future clinical implications”. The difference comparing results of AI and manual approaches (applying a receiver operating characteristic[ROC] analysis) reported an increase of 18.7% (ROC measurement of 0.991 with AI, versus 0.835 with manual).
While Artificial intelligence (AI) and “machine learning” (ML) are often used interchangeably, AI specifically refers to the “general ability of computers to emulate human thought and perform tasks in real-world environments”, and machine learning refers to the “technologies and algorithms that enable systems to identify patterns, make decisions, and improve themselves through experience and data”. The field of AI intersecting on healthcare is exploding and there appears to be no end of applications on the horizon. According to Frost & Sullivan (a US business consulting firm), by 2030, over 50% of jobs in the Western world could be automated by AI and that around 80% of the healthcare business models in 2025 will be driven by platform-based healthcare analytics and intelligence solutions. Peer-reviewed experts in the medical field suggest that AI may soon replace certain aspects of work from radiologists, cardiologists and ophthalmologists, perish the thought! However, the current study has aimed to investigate the efficacy of “AI to generate multi-instrument structural and functional imaging studies” by overlaying images from a cohort of patients diagnosed with RP.
The recent US study used infrared images from microperimetry on near-infra-red images from SLO and SD OCT in RP patients using manual alignment and AI. Manual alignment was performed using in-house software that allowed labelling of six key points located at vessel bifurcations and the manual overlay was considered successful if the distance between same key points on the overlayed images was ≤1/2°. Following the study, fifty-seven eyes of 32 patients were included in the analysis showing that AI was significantly more accurate and successful in aligning images compared to manual alignment, as confirmed by linear mixed-effects modelling (p < 0.001). A receiver operating characteristic analysis, used to compute the area under the curve of the AI (0.991) and manual (0.835) (a difference of an improvement of 18.7%). Commenting on the work, the researchers stated that, “using such precision overlays, it will be possible to correlate and monitor functional deficits as evidenced by micro-perimetric retinal sensitivity in small areas with structural changes in retinal degenerations”.